Legal claims defining the scope of protection, as filed with the USPTO.
1. A method of increasing speed of an asynchronous pulse processing based hyperspectral target detection algorithm, the method comprising: using stored optimized filter coefficients to provide initial filter coefficients; optimizing initial filter coefficients using an asynchronous pulse processor based hyperspectral detection algorithm to provide optimized filter coefficients, wherein the asynchronous pulse processor comprises: an initialization module configured to set an instantaneous constrained energy minimization (CEM) filter coefficients vector to comprise a plurality of stored optimized filter coefficients; a pixel-filter coefficient matrix product module configured to calculate a CEM value vector based on the instantaneous CEM filter coefficients vector; and a filter coefficient calculation module configured to provide the instantaneous CEM filter coefficients vector based on the CEM value vector, and store the optimized filter coefficients; and storing the optimized filter coefficients.
2. The method according to claim 1 , further comprising running the asynchronous pulse processor based hyperspectral detection algorithm until the optimized filter coefficients converge.
3. The method according to claim 1 , further comprising running the asynchronous pulse processor based hyperspectral detection algorithm until a constrained energy minimization (CEM) value becomes confined within a bound for a pre-determined time.
4. The method according to claim 1 , wherein: the asynchronous pulse processor further comprises a background spectra estimation module configured to estimate a background spectra vector; and the filter coefficient calculation module is further configured to provide the instantaneous CEM filter coefficients vector based on the background spectra vector.
5. The method according to claim 1 , wherein: the asynchronous pulse processor further comprises a target-filter coefficient matrix product module configured to calculate a CEM value result; and the filter coefficient calculation module is further configured to provide the instantaneous CEM filter coefficients vector based on the CEM value result.
6. The method according to claim 1 , wherein the hyperspectral detection algorithm comprises a constrained linear programming optimization.
7. A system for increasing speed of an asynchronous pulse processing based hyperspectral target detection algorithm, the system comprising: an asynchronous pulse processor configured to: set initial filter coefficients to stored optimized filter coefficients; optimize the initial filter coefficients using a hyperspectral detection algorithm to provide optimized filter coefficients; and store the optimized filter coefficients, wherein the asynchronous pulse processor comprises: an initialization module configured to set an instantaneous constrained energy minimization (CEM) filter coefficients vector to comprise a plurality of stored optimized filter coefficients; a pixel-filter coefficient matrix product module configured to calculate a CEM value vector based on the instantaneous CEM filter coefficients vector; and a filter coefficient calculation module configured to provide the instantaneous CEM filter coefficients vector based on the CEM value vector, and store the optimized filter coefficients.
8. The system for target detection from hyper-spectral image data according to claim 7 , wherein the hyperspectral detection algorithm comprises a constrained linear programming optimization.
9. The system for target detection from hyper-spectral image data according to claim 7 , wherein the asynchronous pulse processor is further configured to run until the optimized filter coefficients converge.
10. The system for target detection from hyper-spectral image data according to claim 7 , wherein the asynchronous pulse processor is further configured to run until a constrained energy minimization (CEM) value becomes confined within a bound for a pre-determined time.
11. The system for target detection from hyper-spectral image data according to claim 7 , wherein: the asynchronous pulse processor further comprises a background spectra estimation module configured to estimate a background spectra vector; and the filter coefficient calculation module is further configured to provide the instantaneous CEM filter coefficients vector based on the background spectra vector.
12. The system for target detection from hyper-spectral image data according to claim 7 , wherein: the asynchronous pulse processor further comprises a target-filter coefficient matrix product module configured to calculate a CEM value result; and the filter coefficient calculation module is further configured to provide the instantaneous CEM filter coefficients vector based on the CEM value result.
13. A circuit for asynchronous pulse processing based hyper-spectral target detection, the circuit comprising: an initialization module configured to set an instantaneous constrained energy minimization (CEM) filter coefficients vector to comprise a plurality of stored optimized filter coefficients; a pixel-filter coefficient matrix product module configured to calculate a CEM value vector based on the instantaneous CEM filter coefficients vector; and a filter coefficient calculation module configured to provide the instantaneous CEM filter coefficients vector based on the CEM value vector, and store the stored optimized filter coefficients, wherein the circuit is configured to: set initial filter coefficients to stored optimized filter coefficients; optimize the initial filter coefficients using a hyperspectral detection algorithm to provide optimized filter coefficients; and store the optimized filter coefficients.
14. The circuit according to claim 13 , wherein the pixel-filter coefficient matrix product module and the filter coefficient calculation module are further configured to run until the optimized filter coefficients converge.
15. The circuit according to claim 13 , wherein the pixel-filter coefficient matrix product module and the filter coefficient calculation module are further configured to run until at least one of CEM value of the CEM value vector becomes confined within a bound for a pre-determined time.
16. The circuit according to claim 13 , further comprising a background spectra estimation module configured to estimate a background spectra vector, wherein the filter coefficient calculation module is further configured to provide the instantaneous CEM filter coefficients vector based on the background spectra vector.
17. The circuit according to claim 13 , further comprising a target-filter coefficient matrix product module configured to calculate a CEM value result, wherein the filter coefficient calculation module is further configured to provide the instantaneous CEM filter coefficients vector based on the CEM value result.
18. The circuit according to claim 13 , wherein the circuit is configured to perform a constrained linear programming optimization.
Unknown
June 17, 2014
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.